Heats of Formation of Solid Binary Compounds. 35 



points is 11 per cent., whilst the average difference between 

 the heats of formation as found experimentally for the same 

 compound by different observers is 8 per cent.; which shows 

 that the method of calculating the heats of formation from the 

 melting-points gives results as near as could be expected, and 

 especially when we take into consideration : — 



(1) That no less than three separate melting-points are 

 employed in each calculation, thus rendering the latter liable 

 to three sources or error. 



(2) That the size of the molecules in the solid state, both of 

 the elements entering into the reaction and of the compounds 

 produced, are in most cases quite unknown. 



(3) That many of the melting-points employed are subject 

 to considerable uncertainty; thus those for Ba vary between 

 450° C. and the melting-point of cast iron, those for Cu be- 

 tween 950° and 1330°, those for Ag between 916° and 1040°, 

 for Mg between 433° and 750°, those for Al between 600° and 

 above 890°, for Cd between 228° and 325°, for Zn between 

 360° and 450°, for Pb between 320° and 335°, for Sn between 

 220° and 246°, and for Sb between 425° to 621°. 



(4) That the high average difference between the calculated 

 and experimental heats of formation is chiefly due to large 

 errors in a comparatively few cases ; for only about one third 

 (14 out of 38) of the whole number of instances give differ- 

 ences above the average, and in 6 out of the 14 only one expe- 

 rimental determination of the heat of formation has been made, 

 whilst in four of the remaining 8 the difference between the 

 experimental numbers themselves is very great. Further, in 

 8 of the 14 instances in which differences above the average 

 occur, the determinations of the melting-points of the metals 

 differ by as much as 100° to 400° C, and in the case of Ea 

 even about 1200°. 



On the other hand, in some cases it is only by adopting 



molecules of a size different from that usually ackowledged 



that numbers agreeing with our theory are obtained. Also in 



N 

 calculating the ratio ^ we have in a few instances used that 



heat of formation most favourable to the theory. 



In the tables we have given all the examples to which our 

 method of calculation can be applied ; but whether the above 

 relation does exist or not requires confirmation in other 

 cases, for which, however, the necessary data are at present 

 wanting. 



D2 



